A systematic review published in 2017 put adult lactose malabsorption at about 68 per cent worldwide, using data from 62,910 participants in 89 countries. The journal that published it, Lancet Gastroenterology & Hepatology, retracted the paper in January 2025.
The review, by Storhaug, Fosse and Fadnes, appeared in the October 2017 issue, on pages 738 to 746. Its PubMed record carries a retracted article label, as checked on 20 November 2026. Its regional estimates ranged from 28 per cent to 70 per cent.
The journal published an expression of concern in its September 2024 issue, at page 784 (PMID 39127069). The retraction followed in the January 2025 issue, at page 13 (PMID 39674208), after the editors found that country estimates relied on studies not representative of the general population or failing other eligibility criteria.
What the review counted
The authors searched MEDLINE and Embase from inception to 2 November 2016 and screened 2,665 records. Nothing published after that search date could enter the dataset. The review covered adults and children aged 10 years or older, so it was not an estimate for the whole population. Children younger than 10, self-reported data, and inpatients and outpatients at gastroenterological wards were excluded. The protocol was registered with PROSPERO as CRD42017064802.
The analysis rested on 450 study populations. That figure counts populations, not individual people. Of those, 306 came from 116 full-text articles used as primary sources. Data for a further 144 study populations were taken from 59 review articles, because the full-text primary articles could not be obtained.
That means 144 of the 450 study populations reached the analysis through review articles rather than through the original reports. Together the 450 populations covered 62,910 participants from 89 countries. The paper said that was 84 per cent of the world's population.
The testing methods were mixed. Genotyping was the most used, in 231 of the 450 study populations. Lactose tolerance tests were used in 101, hydrogen breath tests in 83, and other methods or methods not described sufficiently in 35.
Malabsorption is not the same as intolerance
The 68 per cent figure describes lactose malabsorption, assessed by genotyping, hydrogen breath tests or lactose tolerance tests. It is not a count of people who report symptoms after consuming lactose. Self-reported data were excluded from the review, so symptom reports sat outside its scope.
The country-size-standardised global estimate was 68 per cent, with a 95 per cent confidence interval of 64 to 72. Regional estimates ran from 28 per cent (19 to 37) in western, southern and northern Europe to 70 per cent (57 to 83) in the Middle East.
Both regional figures carry wide confidence intervals, 19 to 37 per cent for the European grouping and 57 to 83 per cent for the Middle East.
The review also reported a separate pooled estimate for each test type, and the estimate moved with the test used. Genotyping alone gave 74 per cent (69 to 80), the highest of the three. Lactose tolerance tests gave 55 per cent (46 to 65), the lowest. Hydrogen breath tests gave 57 per cent (46 to 67).
The abstract recorded variation inside countries as well. "There was substantial heterogeneity between studies within most of the assessed countries," it said.
The authors' summary at the time read: "Lactose malabsorption is widespread in most of the world, with wide variation between different regions and an overall frequency of around two-thirds of the world's population." That interpretation now sits under a retracted paper.
Why the journal retracted the paper
In the September 2024 expression of concern, the editors wrote that concerns had been raised in March 2024 about the inclusion of studies not representative of the general population, which was one of the review's stated exclusion criteria. They reported that the authors confirmed the Czech data came from a subpopulation of Roma people and that the data for Oman and Yemen might not represent the general population.
In the January 2025 retraction notice, the editors said they had worked with the authors to examine further data and identified estimates for a substantial number of countries derived wholly or partly from studies not representative of the general population or failing other eligibility criteria. Concluding that the article's findings are unreliable, the journal retracted the paper.
The paper appeared in October 2017. The expression of concern followed in September 2024 and the retraction in January 2025.
PubMed now carries all three records, with the 2017 article under a retracted article label. The article is indexed at PMID 28690131, the expression of concern at PMID 39127069 and the retraction at PMID 39674208.
The figures above are what the authors reported in 2017, in a paper the journal has since retracted.
For VegOut coverage of dairy alternatives and soy milk, see how precision fermentation makes whey protein without a cow and how much global soy goes to tofu and soy milk versus livestock feed.
Photo: A glass of milk. Santeri Viinamäki, CC BY-SA 4.0, via Wikimedia Commons. Share-alike licence.




